[GSoC-2009] Shallow Clones

[Matt Mackall]

On Wed, 2009-04-01 at 20:54 +0530, Madhusudan C.S wrote:
> Hi all,
>      This is a rough draft proposal I have come up with for Shallow
> clones based on the Plans that Peter just sent in to the list. I
> request you all to review it for me. Since I am running out of time
> and djc urged all students to submit it to Google by today, I am
> submitting it there too. I continue editing it and making changes as
> per community suggestions and opinion. (Please point out if there are
> any kind of mistakes, including technical, spelling and grammatical.
> Since this is the first version, I am doing all the work to revise the
> proposal to correct such mistakes, your suggestions will help me make
> it faster.)

Given the complexity and number of possible design trade-offs here,
I consider this particular project to still be in the research stage.

In some ways the most important part of this whole project will be
engaging with the community to settle on a good design. That process
will likely not happen 'all on paper' - it's going to need some
experimentation, possibly multiple proofs-of-concept, and lots of
discussion. It may in fact take the bulk of the GSoC time, with a
successful GSoc result being a prototype that's a solid foundation for a
complete feature.

On the other hand, it's also quite possible that we'll settle on a good
path early on (maybe Peter will convince me that his is the right one).

My point is: the schedule and the breakdown of the current design is not
very interesting to me - it might all be irrelevant next week. It's much
more important to me that you demonstrate understanding of the
complexities of the problem. A detailed breakdown of the problem with
attention to the 'traps' of particular approaches and some weighing of
the trade-offs would be much more useful.

-- 
http://selenic.com : development and support for Mercurial and Linux